IFR R Code

## Affan's reproduce

library(distributions3)
library(ggplot2)
library(tidyverse)
library(R.utils)
library(zoo)
library(tictoc)

#load("formatted_US_death_data.rds")
#US.deaths<-US.deaths %>% filter(DailyDeaths>=0)

# Seyed's data now
US.deaths = fread("./USData.csv")
US.deaths$DailyDeaths = as.numeric(US.deaths$DailyDeaths)
US.deaths$Date = as.Date(US.deaths$Date)

#Best-Guess Estimate for Infection Fatality Ratio (Derived from Lancet Paper)
Cases_Distr <- function(Deaths){
  ifr = rlnorm(1,meanlog = log(0.2), 0.25)
  ifr2 = rlnorm(1,meanlog = log(2.58), 0.212)
  
  #Deaths$Infections <- Deaths$DailyDeaths/ifr*100  ###Move this to the end
  df = data.frame(InfectedDate=NA,Country=NA,Count=NA)
  for (i in 1:nrow(Deaths)){
    x<-Deaths[i,]
    #print(x$Infections)
    lags <- rlnorm(x$DailyDeaths, meanlog = log(13.2), sdlog = 0.24)
    incubationperiods = rlnorm(x$DailyDeaths, 1.434, 0.661)
    lags = lags + incubationperiods
    lagdates <- x$Date - lags
    if (length(lagdates)==0){
      next
    } else{
      s <- data.frame(Count=summary(as.factor(as.character(lagdates))))
      s$InfectedDate <-rownames(s)
      s$Country=x$Country
      s<-s[c(2,3,1)]
      df <- rbind(df,s)
    }
  }
  df <- df %>% group_by(InfectedDate,Country) %>% summarize(Infected = sum(Count)) %>% arrange(InfectedDate) %>% 
    filter(InfectedDate!="(Other)")
  #print(df[1:10,])
  df$InfectedDate<-as.Date(df$InfectedDate)
  df$Country<-as.factor(df$Country)
  df$Infected = 0.31*df$Infected/(ifr/100) + 0.69*df$Infected/(ifr2/100)
  return(df)
}

 
#Deaths = data.frame (US.deaths) 
# add country
MCRuns <- function(Deaths,n){
  trials <- data.frame(InfectedDate=NA,Country=NA,Infected=NA,run=NA)
  pb <- txtProgressBar(min = 0, max = n, style = 3)
  for (i in seq(1,n,1)){
    setTxtProgressBar(pb, i)
    draw <- Cases_Distr(Deaths)
    draw$run <- i
    trials <- bind_rows(trials,draw)
    names(trials)<-c("InfectedDate","Country","Infected","run")
  }
  return(trials)
}

tic()
lancet <- MCRuns(US.deaths, n=30)
toc()

lln2 = lancet %>% group_by(InfectedDate) %>% summarise(infmean = mean(Infected))

plot(lln2$infmean)
sum(na.omit(lln2$infmean))